A HYBRID
ANT COLONY OPTIMIZATION
ALGORITHM
FOR
SOLVING ROUTING PROBLEM IN WIRELESS SENSOR
NETWORKS
By
Mengtao Ji
A thesis proposal submitted to the
Faculty of Business
in
partial fulfillment
of the requirements for the degree of
Bachelor of Information System with Honors in Enterprise System
Gerald Schwartz School of Business
St. Francis Xavier University
Antigonish, Nova Scotia
Mengtao Ji
April 8, 2011
ABSTRACT
TABLE OF CONTENT
ABSTRACT
................................
................................
................................
................................
......
2
TABLE OF CONTENT
................................
................................
................................
....................
3
CHAPTER
Ⅰ

INTRODUCTION
................................
................................
................................
.....
5
1.1
Background
................................
................................
................................
................................
.
5
1.2
Purpose
................................
................................
................................
................................
.......
6
1.3
Research Question
................................
................................
................................
......................
7
1.4
Operational Definitions
................................
................................
................................
..............
8
1.5
Conclusion
................................
................................
................................
................................
10
CHAPTER
Ⅱ

LITERATURE REVIEW
................................
................................
......................
11
2.1
Introduction
................................
................................
................................
..............................
11
2.2
Challenge
................................
................................
................................
................................
..
12
2.3
Research Trends
................................
................................
................................
........................
14
2.4
Related Protocols
................................
................................
........
Error! Bookmark not defined.
2.5
Ant Colony Optimization
................................
................................
................................
..........
12
2.6
Conclusion
................................
................................
................................
................................
18
CHAPTER
Ⅲ

RESEARCH OBJECTIVES
................................
................................
.................
19
3.1
Introduction
................................
................................
................................
..............................
19
3.2
Research and Research Questions
................................
................................
............................
19
3.3
Assumptions
................................
................................
................................
..............................
22
3.4
Conclus
ion
................................
................................
................................
................................
22
CHAPTER
Ⅳ

RESEARCH METHOD
................................
................................
........................
23
4.1
Introduction
................................
................................
................................
..............................
23
4.2
Secondary Data Sources
................................
................................
................................
...........
23
4.3
Conclusion of Method
................................
................................
................................
...............
23
CHAPTER
Ⅴ

PROPOSED DATA ANALYSIS
................................
................................
..............
25
5.1
Parameter Analysis
................................
................................
................................
...................
25
5.2
Convergence Analysis
................................
................................
................................
...............
25
5.3
Validity Analysis
................................
................................
................................
........................
25
CHAPTER
Ⅵ

CONCLUSION
................................
................................
................................
......
26
6.1
In Summary
................................
................................
................................
...............................
26
6.2
Limitations and Future Research
................................
................................
..............................
26
CHAPTER
Ⅶ

PROPOSED TIMELINE
................................
................................
......................
27
7.1
Overall Required Time
................................
................................
................................
..............
27
7.2
Timeline
................................
................................
................................
................................
....
27
REFERENCES
................................
................................
................................
...............................
28
CHAPTER
Ⅰ
IN呒佄UC呉低
1.1
Background
Wireless Sensor Networks (WSNs) are a new
kind of wireless networks that
are
composed of a large number of sensor nodes for the purpose of collect
ing and
transmitting
information
to end

users
from a
sensor field (Naka
mura, Loureiro, &
Frery, 2007).
This new technology is
receiving increased
focus because of the
combination of
wireless communication and
sensor techniques. Many
physical
parameters
in the surroundings
like temperature, humidity
, acou
stic
vibration
,
pressure
,
and
electromagnetism
can be detected by d
ifferent kinds of sensor nodes
.
F
or those various ki
nds of nodes and the
ir
communication abilities,
WSN
can b
e used
for
many
applications such as disaster relief, environment
al
control
(e.g. for detecting
seismic disturbance)
, precision
agriculture
, medicine and health care
(
Karl
&
Willig,
2005
)
.
For example, environment
al
problem
s such as wild fires are
becoming
big
issue
s
.
An efficient
way to
prevent the wild fires
is
important.
N
onetheless
,
wild area
s
like forest
s
are
hard to enter
to record
accurate data
, let alone protect
and
monitor
in
rea
l
time
. Sensor nodes are able to collect
the surrounding
physical
information
and
transmit
the data to the end

users
by
communicating with the
neighbour
nodes
even
in
the forest.
There are so many applications for
WSNs
that they are too numerous to
mention by name.
C
onstructing a proper and efficient
application
for
WSNs
is
difficult
because the
application
requirements
for specific
areas
a
re di
fferent, and
there are also
s
ome
intrinsic limitation
s
for
sensor nodes
like
low power
,
limited lifetime
,
and
maintainability
. In order to construct a
specific application,
researche
r
s find different
ways
to solve those problems such as improving
data processing
abilities
for
nodes,
using efficient routing protocols
for networks
,
and reducing the energy consumption
when nodes are inactive.
1.2 Purpose
This research focuses on
the routing problem
,
which
relates
to
choosing
which
neighbour
nodes
to use
for
transmit
ting
the dat
a
based on network layer
.
Karl
and
Willig
(2005)
concluded
that
“
the constr
uction and maintenance of the
routing tables
which list
the most appropriate
neighbour for any given data destination are used
,
is
the crucial task of
a distributed routing protocol
”
(p.289).
There are many kinds of
routing protocols using different algorithms, and each of them
has a
specific purpose.
This research
will
find a
n
energy efficiency
path in the
WSNs
using
a method is based
on
Ant Colony
Opt
imization
(ACO)
algorithm
s
,
aiming to
minimize the consumption
of power, improve fault tolerance
,
and lengthen the network
lifetime
.
Ant
C
olony
O
ptimization
is one kind of algorithm
which is inspired
by
swarm intelligence
(
Dorigo
&
Stützle
,
2004
). An ant
will leave pheromones for other ants when finding
the food. Some ants will follow
the pheromone to find the food,
but some other ants
will find a shorter path and also deposit their
pheromone. Eventually, the shortest path
from nest to food source will be
found
.
Inspired
by
the activity of ants in nature, the
prototype of the
ACO System
was
built
by
Dorigo
to solve intelligence optimization
problems
.
Since then, Dorigo
and other researchers
have been working
on ACO
system and extend
ing the system. They have
already proposed Ant Colony System
(ACS)
and
MAX

MIN Ant System (MMAS)
which
both are
metaheuristic
algorithms
(
Dorigo
& Socha,
2007
).
ACO algorithms focus on
global
search
, but their local
search is not so efficient
. This
inefficiency
may cause
increas
ed
consumption of power if there are a
large number
of
paths after th
e first global search. Compar
ed
to
other
algorithms, single ACO may
have a
poor
performance when the number of nodes is large.
Lim, Jain
and
Dehuri
(2009
) have already
found
that
ACO
with local search can
perform well on
some
complex problems.
Therefore,
an improved
A
nt
C
olony
O
ptimization
which
combined with
local search
is
proposed
for s
olving this complex optimization
problem
.
Th
ere are two
general methods of local search
:
genetic
local search and
simulated annealing
.
T
he
results of
two
hybrid
Ant Colony
Optimization a
lgorithms
will be compared to the
original
ACO algorithm and other routing results,
to
determine whether ACO combined with local search is more effective or not.
1.3 R
esearch Question
Because
of the
great application value of
WSNs
and the success of
ACO
, there is a
need to
solve the routing problem adapting for the specific area.
The major research
que
stion is how to find a
n energy efficiency
path in
WSNs
using
a hybrid
ACO
a
lgorithm. As previously mentioned,
answering
this research question
will improve
the ability to construct
a power efficient Wireless Sensor Network
for a specific
situation
. Also, the fault

tolerance and network lifetime will be covered as th
e
measures for analyzing the result. For these
purpose
s
, four sub

questions can be
divided according to the processes sequence.
First,
w
hat are the features of routing
problem
s
in
WSNs
?
This question helps to
understand the existing protocols
, which
are
fo
r the following model building. Second
,
h
ow
can
a mathematical model of
routing pro
blem
be constructed
using ant colony algorithm
and
two
hybrid ACO
algorithm
s
?
This research will construct three
model
s
of
WSNs
which are
trying to
transform a real

life problem to a structured math problem
. The hybrid algorithm
models are
based on the fundame
ntal ACO algorithm model, but they all
reinforce
local search. Building
a
math model can easily explain and analyze
the problem.
Thir
d
,
how can a simulation of those
mathematical model
s
be
design
ed
?
This question
will
develop a simulation by some related computer software, so this question is
important to produce the result. Fourth,
how can the result be analyzed
?
The result of
this res
earch will cover the final path
s from the three
models
and other related
parameters. The conclusion will be obtained by analyzing the data.
1.4
Operational Definitions
Wireless Sensor Networks (WSNs) are a new
kind of wireless networks that
are
composed
of a large number of sensor nodes for the purpose of collect
ing and
transmitting
information
to end

users
from a
sensor field (Nakamura, Loureiro, &
Frery, 2007).
Ant Colony Optimization (ACO) is an algorithm used for solving the shortest path
problem which is inspired
by real
ants and inclu
ded in metaheuristic algorithms
(
Dorigo
&
Stützle
,
2004
).
A sink node is a node which collects data from all the sensor nodes
in the sensor field
and sends the data to the users
(
Karl
&
Willig, 2005
)
.
Pheromone is an excreted chemical factor among the same species. It can be
deposited on the ground when ants look for food. The other ants will receive the food
information from ph
eromone on the ground
(
Dorigo
&
Stützle
,
2004
).
Constraint energy problem is a main issue in Wireless Sensor Network. It is resulted
from low power storage in batteries, using energy inefficie
ntly and bad routings in
WSNs
(
Karl
&
Willig, 2005
)
.
Genetic local search is the combination of Genetic Algorithm (GA) and local search
heuristic
and GA is an algorithm which is inspired from genetic operations (
Gen &
Cheng,
2000
).
Simulated Annealing is one kind of local search algorithms to accept thresho
ld in a
randomized version (
Mich
iels, Korst,
& Aarts,
2007
).
1.5
Conclusion
The second chapter will review some relevant liter
at
ures,
and give the major gaps of
them. The third chapter will state the research question and sub

questions
, and give
more deta
ils of their objectives. The fourth chapter will state the research methods.
The fifth methods will mainly discuss the plan to analyze the data.
CHAPTER
Ⅱ
䱉呅TA呕R䔠E䕖I䕗
2.1
Introduction
T
he
literature of a synthetical review was obtained from the S
tFX Macdonald Library,
ABI/INFORM Global
database, ACM Digital Library
database
, SciVerse
ScienceDirect database
and
American Mathematical Society MathSciNet database
.
All the books and journal articles
are used for better
understanding this
research and
addressing the research q
uestions. The
synthetical
review cover
s the definitions and
characteristics
of
WSNs
, Ant Colony Optimization, the aspects local search,
and some
models of
WSNs
for solving the
constraints
.
Due to the
proprietary
natur
e of the
literature, few references and in

text citation are made for
necessary
.
The procedure of this literature review includes four steps. First is to search for the
definitions and
characteristics
of
WSNs
, and also knowing their challenges.
Second
ly
,
the fundamental principles of Ant Colony Optimization
are introduced from the book
“
Ant Colony Optimization
”
written by Dr. Dorigo and his research members (
2004
)
.
Also, some brief theories of genetic local search and simulated
annealing
will be
covered a
fter understanding ACO algorithms.
Third
ly
, a
ccording to the challenges,
few
popular researches for solving those problems will be
reviewed, and the materials
are mainly covered the differen
t protocols of network routing, so the techniques and
tools in thi
s
research can be decided. F
inally
,
according to the similar profession
al
articles, comparing two
improved
ACO algorithm
which hybrid
s with genetic local
search and simulated
annealing
with
other routing has
not been
researched
before
.
2.2
Wireless Sensor
Networks
According to the article of
Nakamura, Loureiro
and
Frery
(2007
), a clear
concluded
definition
of Wireless Sensor Networks is that
WSNs
is
a ne
w kind wireless network
which are
composed by a l
arge number of sensor nodes
for
the purp
ose of collecti
ng
and transmitting
information from the sensor field
to end

users. Th
e
article presents
the
information fusion problems in
WSNs
, and it also gives a very clear and specific
characteristics and limitations of WSNs nowadays. Information fusion in
WSNs
cover
s many levels of applications including internal tasks (e.g., data routing) and
system applications (e.g., target detection). Also, many methods, techniques and
algorithms
for solving different layer problems
are
in
troduced
in t
he
article.
Communication la
yer includes routing
problems
, localized problems and so on, and
many
protocols
are raised to solve routing problems, each of them directs at a specific
application.
In addition,
Karl
and Willig
(2005)
mainly introduces the network architecture and
communication protocols in WSNs
. Communication protocols includes physical layer,
MAC protocols, naming and addressing, time
synchronization
, localization and
positioning, topology control, routing protocols, data

centric and content

based
networking, tran
sport layer and quality of
service and advance application
.
Due to this
research
will not
cover
deployment environment, data interface and nodes position, so
the major focuses are topology control and
routing protocols
.
Karl
and Willig
(2005)
presented tha
t there are
three options to design topology which are flat networks,
hierarchical networks with backbones and hierarchical networks with clusters.
According to the three options of topology models, t
his
research
will use
flat networks
which are the some e
nergy efficient links of all the active nodes can be linked instead
of all links in an original density
networks.
Also,
Karl
and Willig
(2005)
introduce
d
diff
erent protocols and algorithms
,
and
compared
the various protocols to understand
the specific purp
oses of the protocols.
2.3 Ant Colony Optimization
The book
“
Ant Colony Optimization
”
written by
Dorigo
and
Stützle
(2004)
is a good
introduction
material. In the nature, ants use a chemical substance called pheromones
to find food for giving hints to other ants. Other
ants follow the pheromone to
get to
the food
. If there are two ways to get to the
food source
,
a short one and a long one,
and ther
e are a group of ants randomly choosing two ways to get to the destination.
The ants
that
choose the short path will first arrive at the destination and go back to
their nest, so there are stronger
pheromones
on the short path. This is a famous
experiment
called double

bridge experiment. Double

bridge experiment demonstrates
that ants can find
the shortest path from nest to food
source
.
Inspired from
the activity of ants in nature, the prototype of the
ACO System was
built to solve intelligence optimizati
on
problems.
Since then, Dorigo, Gambardella,
and
Stützle
have been working on ACO system and extending the system.
Dorigo and
Socha (2007) published an article,
“
An introduction of Ant Colony Optimization
”
,
and it gives a brief introduction
to the audienc
es who are first touch this algorithm.
The article shows that different kinds of ACO algorithms like Ant
Colony System
(ACS) and MAX

MIN Ant System (MMAS) which both are
metaheuristic
algorithms
.
Many
extension
and improvement of ACO are presented to optim
ize the parameters or
to solve some combinational problems. For the combinational optimization problems,
ACO algorithms
can be
hy
b
rid
with the algorithms which is good at local search.
In addition, the book
“
Theoretical Aspects of Local Search
”
written by
Michiels
,
Korst
and
Aarts
(2007
) presented several methods to solve local search optimization,
which includes genetic local search and
simulated
annealing. These two algorithms
will be used for improving a performance of energy efficient.
Also,
“Computational
Complexity of Ant Colony Optimization and its Hybridization with Local Search”
(
Lim,
Jain, & Dehuri, 2009
) indicated that the hybridization of ACO with local search
will obtain a
successful
effect for some complex problems, but the paramete
rs of ACO
algorithm will be the important factor for the final results.
2.4
Re
lated Researches and Models
According to the principles and characteristics of Ant Colony
Optimization
and
Wireless Sensor Networks, many models and protocols of
routing problem
s are
presented aiming to solve the limited lifetime a
nd energy consumption problems.
Some models considered routing problems but not using Ant Colony Optimization
algorithms. For example,
“
Adaptive design optimization of wireless sensor networks
using ge
netic algorithms
”
uses genetic algorithm to solve the energy constraint
problem. This proposal of using genetic algorithm is suitable for all
application

specific requirements and can solve the communication and energy
constraint problem. This research max
imize the lifetime of sensor nodes for all
application

specific requirements
(
Ferentinos
&
Tsiligiridis,
2007
).
The optimization
problem of
Wireless Sensor Networks
in th
e
paper
,
“
Coverage

Time Optimization for
Clustered Wireless Sensor Networks: A
Power

Balancing Approach
”
,
considers the
power balanced coverage time for clustered WSNs. The advantage of clustering is
decreasing energy consumption and the disadvantage is increasing cluster heads’
communication burden. Deterministic setups and stochast
ic setups are the two
methods to solve the optimization problems. Stochastic setups can be divided into two
parts which are a routing

aware optimal cluster planning and a clustering

aware
optimal random relay and they can balance power consumption
(
Shu
&
K
runz
,
2010)
.
Using supervised learning techniques to deal with
the
challenges
is proposed in
“
Optimization for Schemes for Wireless Sensor Network Localization
”
. A supervised
learning framework can produce useful information and make decisions in sensor
ne
twork. It can be applied to other problems which could benefit from information
discovering. This research provides a new direction to routing optimizations for
deployi
ng the wireless sensor networks
(
Szynkiewicz
& Marks, 2009
).
In addition, some article
s include routing problems but they also cover other layer
problems, like
“
Modeling and Optimization of Transmission Schemes in
Energy

Constrained Wireless Sensor Networks
”
.
This paper models three layers
which are the link layer, the medium access control
layer and the routing layer to
identify the traditional network. At first, the assumption of computing a maximum
network lifetime is based on one layer at a time and the other layers are
fixed.
T
his
paper constructs that the optimization problem could be
solved by cross

layer of Time
Division Multiple Access (TDMA)
,
and it also gives some examples to illustrate the
b
enefits of cross

layer designin
g (
Madan,
Cui,
Lall,
& Goldsmith,
2007
)
.
Some models used Ant Colony Optimization or Ant Colony Optimization e
xtension to
solve routing problems. For example,
“A Novel Routing Protoc
o
l in Wireless Sensor
Networks based on Ant Colony Op
t
imization
”
is going to
Maximiz
e
the coverage
time by balancing the power consumption of different cluster heads (CHs) is a major
optimization problem of
Wireless Sensor Networks
. Two mechanisms are proposed
for achieving balanced power consumption. One is the routing

aware optimal clust
er
planning and the other one is the clustering

aware optimal random relay. The research
demonstrates that both mechanisms maximize the power consumption of CHs. The
optimization problems are formulated as signomial optimizations and they are solved
effici
ently. This research demonstrates that the two schemes substantially prolong the
coverage time of the network
(
Xie,
Zhang,
&Feng,
2010).
“
Parallel energy

efficient
coverage optimization with maximum entropy clustering in wireless sensor networks
”
construct
s a parallel energy

efficient coverage optimization mechanism with
maximum entropy clustering to solve energy constraint problem in w
ireless sensor
networks.
A
ll the sensor nodes are divided into clusters by maximum entropy
cluste
ring
, and
Dijkstra’s algor
ithm is used to calculate the lowest cost paths inside
each
cluster.
Then
swarm optimization is used to solve the problem of maximizing the
coverage metric and minimizing the energy metric for parallel clusters. A trade

off
between coverage rate and energy
effici
ent can be solved in this paper
(
W
ang, Ma, &
Wang,
2009
).
2.5 Gaps
According to those
researches
and models on
the
routing problems, all of them are
going to solve
energy constraint
and limited lifetime
problem
s. Besides those two
problems, there a
re
many other challenges like
maintainability
, fault tolerance, density
controlling,
Quality of Service (QoS) and so on (
Karl
& Willig,
2005
). All of those
are the major
challenges
in WSNs,
and
the associated requirements includes long
lifetime, low energy cost, self

adaption (some broken nodes will not
influence
the
entire network),
small density, good quality of service and so on. A
lot of researchers
used different algorithms trying to op
timiz
e
as many requirements as possible.
In this
researc
h, the major objective is
comparing two
improved ACO algorithm which
hybrids with
genetic local search and simulated
annealing
to solve the routing
problem in WSNs.
W
ang, Ma
and
Wang
(
2009
)
constructs a
parallel energy

efficient coverage
optimization mechanism with maximum entropy clustering to solv
e energy constraint
problem
using ACO algorithm which hybrids with Dijkstra algorithm, and it is similar
to this proposed research.
However, t
his research
will
use
two hybrid
ACO
algorithms;
each of them is with genetic local search or
simulated annealing,
instead of Dijkstra
algorithm. Also, Wang, Ma and Wang used the
hierarchical
networks with clusters,
but this research will propose a flat networks. This rese
arch will consider both energy
consumption and fault tolerance to obtain a long lifetime, low energy cost and
self

adaption Wireless Sensor Network.
2.6
Conclusion
In order to solve routing problem in
WSNs
using
a hybrid
ant colony algorithm, the
princip
les
and characteristics
o
f Wireless Sensor Networks and
A
nt
C
olony
O
ptimization will be focused on.
Also,
the aspects of local searches
and some specific
and detail
theories
may appear when
starting to do this research step by step.
Despite
a lot of
researches have been done
to solve the routing problem, comparing two
hybrid Ant Colony Optimization algorithm
s
with
two different local searches
which
are genetic local search and simulated annealing has not been proved
. Thus, this is an
opportunity to th
is research.
CHAPTER
Ⅲ
R䕓䕁RC䠠佂䩅C呉V䕓
3.1 Introduction
The major research que
stion is how to find a
n energy efficiency
path in
WSNs
using
a
hybrid
ACO
a
lgorithm. Also, the fault

tolerance and network lifetime will be covered
as the measures for analyzing the result. For these purposes, four sub

questions can be
divided according to the processes sequence.
This chapter will present the four
sub

questions a
nd discuss the objectives for each of those questions.
The first
question is
w
hat the features of routing problem
s
are
in
WSNs
.
This question helps to
understand the existing protocols, which
are
for the following model building.
The
second question is
h
ow
a mathematical model of routing problem
can
be constructed
using ant colony algorithm
and two hybrid ACO algorithms
. Mathematical
models
are
trying to transform a real

life probl
em to a structured math problem
, which can
simplify the complex problem.
Also, t
he hybrid algorithm models are based on the
fu
ndamental ACO algorithm model
.
The t
hird
question is
how a simulation of those
mathematical model
s
can
be
design
ed
.
This question
will
develop a simulation by
some related computer software, so this ques
tion is important to produce the result.
The last question is
how the result
can
be analyzed
.
The conclusion will be obtained
by analyzing the data.
Also, the methodology of analyzing data will be discussed in
detail in Chapter 5.
3.2
Research Questions
3.2.1
Routing in Wireless Sensor Networks
The first question is
w
hat the features of routing problem
s
are in
WSNs.
This question
helps to understand the existing protocols, which
are
for the following model building.
There are many kinds of networks and pr
otocols for WSNs, and the various protocols
are refer to the specific applications.
Many assumptions of this research will be made
when answering this question.
Due to this research will not
cover
deployment
environment, data interface and nodes position,
so the major focuses are topology
control and routing protocols.
Karl
and Willig
(2005) presented that there are
three
options to design topology which are flat networks, hierarchical networks with
backbones and hierarchical networks with clusters.
According to the three options of
topology models, t
his
research
will use flat networks which are the some energy
efficient links of all the active nodes can be linked instead of all links in an original
density networks.
Also,
Karl
and Willig
(2005)
intro
duce
d
different protocols and
algorithms
,
and
compared
the various protocols to understand the specific purposes of
the protocols.
3.2.2
Mathematical Model
The second question is h
ow a mathematical model of routing problem
can be
constructed
using ant colo
ny algorithm
and two hybrid ACO algorithms
.
Mathematical
models are trying to transform a real

life problem to a structured math
problem
, which can simplify the complex problem.
Also, t
he hybrid algorithm models
are based on the fundamental ACO algorithm m
odel.
Second, H
ow to build a mathematical model of routing problem using ant colony
algorithm
and two hybrid ACO algorithms
?
This research will construct three models
of Wireless Sensor Networks which are trying to transform a real

life problem to a
struct
ured math problem. The hybrid algorithm models are based on the fundamental
ACO algorithm model, but they all reinforces local search. Building math model can
easily explain and analyze the problem.
3.2.3
Simulation
The t
hird
question is
how a simulation o
f those
mathematical model
s
can be
design
ed.
This question
will
develop a simulation by some related computer software, so this
question is important to produce the result.
Third, H
ow to design a simulation of the mathematical model
s
?
This question is going
to develop a simulation by some related computer software, so this question is
important to produce the result.
3.2.4
Analyze
The last question is how the result can be analyzed.
The conclusion will be obtained
by analyzing the data
.
Fourth,
How to analyze the result?
The result of this research will cover the final
paths from the three models and other related parameters. The conclusion will be
obtained by analyzing the data.
3.3 Assumptions
The sensor nodes in wireless sensor netwo
rks are stationary.
The sensor nodes in wireless sensor networks are randomly distributed.
Every sensor nodes are the same. Communication capability, computation capability
and energy are the same for every sensor nodes.
The lifetime of the sensor nodes
is determined by the energy.
Barriers in the sensor field will influence the path.
Ant colony optimization is a solution to the routing problem and it saves the energy
consumption and prolongs the lifetime of sensor nodes. However, it cannot solve
probl
ems like the security level issue in Wireless Sensor Networks (WSNs).
3.4 Conclusion
CHAPTER
Ⅳ
R䕓䕁RC䠠H䕔䡏E
4.1
Introduction
4.2
Secondary Data Sources
4.3 Simulation Tool
4.3 Conclusion of Method
This research will
basically
use secondary data sources and Matrix Laboratory
(MATLAB), a numerical computing environment, to solve the above sub

questions.
The ideas about ant colony optimization and the first ACO system were presented by
Dr. Dorigo, so this research will get the a
nt colony algorithm information most from
the papers written by Dr. Dorigo. The article, Ant Colony Optimization
—
Artificial
Ants as a Computational Intelligence Technique, is one of the most recent articles
which was published on IEEE Computational Intelli
gence Magazine. This paper
introduces the Ant Colony Optimization (ACO), the first ACO system which is also
called Ant System (AS) and Ant Colony System (ACS). Each system has a
formula
to
construct a prototype solving the problems in different rules.
It
is similar to the way getting the information on Ant Colony Optimization. The
information on Wireless Sensor Networks will be obtained online. There are many
journal articles which introduce WSNs. This research will use the articles which
mainly present th
e consumption of energy and routing problems. One article is a
chapter which is from the book
Smart environments technologies, protocols, and
applications
. The article introduces some basic
characteristics
and applications about
WSNs (
Cook,
& Das,2005
). In addition, there is another article
Power consumption in
wireless sensor networks
which is presented the energy problems (
Aslam, Farooq, &
Sarwar,
2009
). After knowing
that
information, it is necessary to select some key
points which are useful to thi
s research. In addition, the final result will be analyzed
and compared to other researches which optimized the routing in WSNs using
different algorithms. Some secondary data on optimizing routing in WSNs will be
included in this research such as the simu
lation area of WSN, the number of sensor
nodes and the result data.
Matrix Laboratory (MATLAB) is a useful tool to design and construct an algorithm
simulation. After completing designing the program, the simulation can be
implemented. In order to analyze
the final result, some parameters like the number of
sensor nodes will be
set
as the same values.
Chapter 4: Research Method (state and defend research method, describe the research
instruments and link to the research and investigative questions
CHAPTE
R
Ⅴ
PR佐体䕄⁄A呁⁁NA䱙卉S
5.1
Parameter Analysis
5.2 Convergence
Analysis
5.3 Validity Analysis
Chapter 5: Proposed Data Analysis (how do you plan to analyze the data, introduce
the techniques in detail, justify the analysis, major issues, possible proble
ms and how
they will be addressed)
CHAPTER
Ⅵ
C低C䱕卉低
6.1 In Summary
6.2 Limitations and Future Research
Chapter 6: Conclusion (summary of the proposed study, expected research
contribution and practical implications to managers)
(3 pages)
CHAPTER
Ⅶ
PR佐体䕄⁔䥍䕌䥎
7.1 Overall Required Time
7.2 Timeline
REFERENCES
Karl, H., & Willig, A. (2005). Protocols and architectures for wireless sensor
networks. Hoboken, NJ: Wiley.
Nakamura, E.F., Loureiro, A.F., & Frery, A.C. (2007). Information
Fusion for
Wireless Sensor Networks: Method, Models, and Classifications.
ACM Computing
Survey, 39
(3), 9/1

9/55.
Dorigo, M., & Stützle, T. (2004).
Ant colony optimization
. Cambridge, Mass: MIT
Press.
Dorigo, M., & Socha, K. (2007). An introduction of A
nt Colony Optimization.
Approximation Algorithms and Metaheuristics
(pp.1

29). Boca Raton: CRC Press.
Madan,R.,Cui,S.,Lall,S.,& Goldsmith, A.J.(2007). Modeling and Optimization of
Transmission Schemes in Energy

Constrained Wireless Sensor Networks.
IEEE/AC
M
Transactions on Networking, 15
(6), 1359

1372.
Wang,X., Ma,J., & Wang, S. (2009). Parallel energy

efficient
coverage optimization
with maximum entropy clustering in wireless sensor networks.
Journal
of Parallel and
Distributed Computing, 69,
838

847.
Ferentinos,K.P., & Tsiligiridis,T.A. (2007). Adaptive design optimization of wireless
sensor networks using genetic algorithms.
Computer Networks, 51
, 1031

1051.
Shu,T.,& Krunz,M.(2010).Coverage

Time Optimization for Clustered Wireless Sensor
Networks: A P
ower

Balancing Approach.
IEEE/ACM Transaction on Networking,
18
(1), 202

215.
Xie,H., Zhang,Z.,&Feng,N. (2010)
. A Novel Routing Protoc
o
l in Wireless Sensor
Networks based on Ant Colony Op
t
imization.
International Journal of Intelligent
Information Technolog
y Application, 3
(1), 1

5.
3. Szynkiewicz,E.N.,& Marks,M.(2009). Optimization for Schemes for Wireless
Sensor Network Localization.
Int.J.Appl.Math.Comput.Sci,19
(2), 291

302.
Gen, M., & Cheng, R. (2000).
Genetic algorithms and engineering optimization
.
Ne
w York: Wiley.
Michiels, W., Korst, J., & Aarts, E. H. (2007).
Theoretical Aspects of Local Search
.
Berlin, Heidelberg: Springer

Verlag Berlin Heidelberg.
Lim, C. P., Jain, L. C., & Dehuri, S. (2009). Computational Complexity of Ant Colony
Optimization
and its Hybridization with Local Search.
Innovations in swarm
intelligence
(pp. 91

120). Berlin: Springer.
Σχόλια 0
Συνδεθείτε για να κοινοποιήσετε σχόλιο